F4-220 bad NFS performance
Posted: 28 Oct 2020, 12:37
Guys,
I am using the F4-220 as a NFS datastore for Vsphere ESXI (5.5) - we replaced an old HP Microserver (AMD chipset) with the Terrastore.
It is used to clone Virtual machines as a fallback emergecy in case of Ransomware or other issues.
We have two major issues
1) We can clone small virtual machines up to about 100GB in size without too many issues (performance is bad but i will touch on that in a minute)
2) When we try and clone larger virtual machines 500GB - the web interface becomes unresponsive (time varies but after about 8 hours) and the only way to recover is to use the power button to power the unit off. We have tried the web interface from a number of different machines and browsers, Firefox (latest), CHrome (Latest), Edge (Latest) - all on 64 bit windows - some windows 10, some windows server 2K8 and 2012.
3) When we perform any clone job we are seeing no more than 5MB/s throughput - our unit is attached to a gigabit switch and is connected at gigabit speed with no errors showing on the port stats on the netgear switch.
4) The ESXI server has a dedicated Intel Gigabit NIC for the connection to the terramaster and is in a VLAN just for these two units.
5) The F4 has 4 x WD Green 2TB drives in a RAID 10 config
6) We have just applied the latest patch on the forum - from here
viewtopic.php?f=28&t=711
7) THis seems to have made the speed worse rather than better - we are currently seeing 2MB/s throughput (according to the Web Interface)
We have upgraded the firmware only yesterday and will wait until the weekend to see if the larger servers can clone or not without locking up the unit. Currently testing with some smaller cloned images around 100GB
Any ideas where to go from here ?
Craig
I am using the F4-220 as a NFS datastore for Vsphere ESXI (5.5) - we replaced an old HP Microserver (AMD chipset) with the Terrastore.
It is used to clone Virtual machines as a fallback emergecy in case of Ransomware or other issues.
We have two major issues
1) We can clone small virtual machines up to about 100GB in size without too many issues (performance is bad but i will touch on that in a minute)
2) When we try and clone larger virtual machines 500GB - the web interface becomes unresponsive (time varies but after about 8 hours) and the only way to recover is to use the power button to power the unit off. We have tried the web interface from a number of different machines and browsers, Firefox (latest), CHrome (Latest), Edge (Latest) - all on 64 bit windows - some windows 10, some windows server 2K8 and 2012.
3) When we perform any clone job we are seeing no more than 5MB/s throughput - our unit is attached to a gigabit switch and is connected at gigabit speed with no errors showing on the port stats on the netgear switch.
4) The ESXI server has a dedicated Intel Gigabit NIC for the connection to the terramaster and is in a VLAN just for these two units.
5) The F4 has 4 x WD Green 2TB drives in a RAID 10 config
6) We have just applied the latest patch on the forum - from here
viewtopic.php?f=28&t=711
7) THis seems to have made the speed worse rather than better - we are currently seeing 2MB/s throughput (according to the Web Interface)
We have upgraded the firmware only yesterday and will wait until the weekend to see if the larger servers can clone or not without locking up the unit. Currently testing with some smaller cloned images around 100GB
Any ideas where to go from here ?
Craig